Sec. iii) TO ANCIENT GRAVEYARDS BY THE KURUK-DARYA 733
search proved equally fruitless. So in order to make sure of the quantity of ice required for longer work on the Kuruk-daryà, it became necessary to camp for a day where we had halted, and to send camels back to Yardang-bulak to fetch the additional supply.
At the end of that day's halt Muhammad Baqir rejoined us with the ice-carrying camels, and his experienced guidance next morning brought us, after a three miles' march east-north-east across ground encrusted with light shôr, to where a flood-bed debouches from the belt of Mesas. There we found the eastern spring we had vainly looked for, with a sheet of ice only 20 yards long formed by it in a narrow channel. The absence of adequate vegetation near it explained why Abdurrahim, when ,guiding Làl Singh, had camped so far away from it. From here I let Muhammad Bâgir return to Yardang-bulak to look after the ponies and baggage that we had left behind there. The rest of us steered to the south-east, along the foot of the steep clay cliffs which mark where the glacis of the outermost Kuruk-tâgh hill chain falls off to the riverine plain towards Lou-lan. L5.1 Singh's plane-table tracing clearly indicated that by keeping below the edge of this glacis we were bound to strike the bed of the ` Dry River ' where it makes a big bend to the northward, and farther on to reach in succession the first two burial-grounds he had noticed (Map No. 29. B. 3).
The route thus followed led over a strip of gravel or stony Sai, less than a mile wide where we started and farther on growing still narrower. This afforded easy going and allowed an open view, so far as the haze now lifting would permit, over the dead riverine belt to the south. On our left the edge of the glacis, which seen from a distance had looked like a continuous line of cliffs from loo to 15o feet high, was found to consist for the most part of a close succession of separate but uniformly aligned Mesas. Erosion by water, cutting through the edge of the clay deposits which form the glacis of the hill chain, had evidently been the principal agent in their formation. This glacis edge, as far as I had occasion to follow it along the Kuruk-daryà, bore everywhere a curious resemblance to an ancient coast-line. I t thus vividly reminded me of the one I had seen to the north of the bay of the salt-encrusted Lop sea-bed, stretching from below Kum-kuduk towards Béshtoghrak.3 On our right the living tamarisk-cones, which had stretched down in a fairly wide belt from where our camp of Yaka-yârdang-bulak had stood, gave way, at a distance of about three miles from the spring, to a zone of dead cones.
Shortly after this, low narrow Yardang ridges appeared between the cones, furrowing the bare clay in the manner I had so frequently observed all over the Lou-lan area, right up to the foreshore-like strip of gravel we were following. But the bearing of the )(di-clangs was here nearly from north to south or else NNE. to SSW., and I found the same direction in other places also along this portion of the course of the Kuruk-darya. This distinct difference from the ENE. to WSW. bearing of the Yârdangs in the Lou-Ian area previously explored by me suggests that along this portion of the dried-up riverine belt the force of the winds sweeping down into the Lop Desert basin from the immediately adjoining ranges and plateaus of the Kuruk-tâgh is the prevailing factor in the process of wind-erosion, while farther to the east the winds drawn into the basin by ` aspiration ' from the side of the Pei-shan and the An-hsi—Tun-huang gap are the stronger agents of this process. Wind-erosion had obviously helped here in sculpturing the Mesas of what might be called the ` coast-line '.
After a total march of eleven miles we struck the bed of the ` Dry River ' where it forms a great northerly bend approaching quite close to the gravel ` foreshore '. The bed was over 150 yards wide and lined, like all the Kuruk-dary5. branches in the Lou-lan area, by rows of dead Toghraks, many here still upright. As far as the eye ranged in the hazy atmosphere, it beheld
3 See above, i. pp. 314 sqq.
II 5 B